Partially Visible Lane Detection with Hierarchical Supervision Approach

ABSTRACT

Advanced Driver Assistant System(ADAS) is a critical component of self-driving cars, and lane detection is one of its applications. Convolution Neural Network (CNN) has exhibited its potential to devise state-of-the-art solutions for various computer vision problems and lane detection. However, different scales of lane lines, partial visibility of lane, and the varying resolution of hierarchical features challenge the CNN-based detection methods to localize the lane accurately. In this work, we propose a deep supervision-based model to detect the partially visible lane lines. Hierarchical (Deep) supervision helps the proposed model to trace and classify the lane features at different scales. The simulation of the proposed model gives robust results with partial visibility of lane paint compared to an existing and proven model. The proposed model's accuracy is as high as 92% for partially visible lane and 99% for full visible lane, which stands to the right in its peer work comparison.

KEYWORDS:

• ADAS
• CNN
• Deep supervision
• Lane detection
• Richer convolution features
• Transfer learning

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the authors.

Additional information

Notes on contributors

Hukam Singh Rana

Hukam Singh Rana received the MTech degree in computer science and data processing from Institute Indian Institute of Technologies Kharagpur India. His current research interests are in the self-driving cars, computer vision machine learning, and pattern recognition.

Thipendra P Singh

Thipendra P Singh is currently positioned as professor and head of Department of School of Computer Science, University of Petroleum & Energy Studies, Dehradun, Uttarakhand, India. He holds PhD in computer science from Jamia Millia Islamia University, New Delhi. He possesses 24 years of experience with him. He has been associated with Tata Group and Sharda University, Greater Noida, NCR, India. His research interests include machine intelligence, pattern recognition and development of hybrid intelligent systems. There are more than dozens of publications to his credit in various national and international journals. Dr Singh is a member of various professional bodies including IEEE, IEI, ISTE, IAENG etc. and also on the editorial/reviewer panel of different journals. Email: [email protected]

Kamal Kumar

Kamal Kumar is working as assistant professor in National Institute of Technology, Uttarakhand, India. He has teaching experience of 18 years. He has obtained his bachelor's and master's from Kurukshetra University, India. He received his PhD from Thapar University, India. He has 35 publications in journals and international conferences. He has served as technical program chair in NGCT'2017, Program Committee members in multiple international conferences. He chaired an International Conference, NGCT'2018. His research interests include in wireless sensor networks, security provisioning, cloud computing and artificial intelligence. Email: [email protected]

Krishan Kumar

Krishan Kumar is serving as an assistant professor in the Department of Computer Science and Engineering at National Institute of Technology Uttarakhand, Srinagar (Garhwal), Uttarakhand, India since July 2014. Dr Kumar received PhD for his thesis ‘Performance Enhancement of Events Detection and Summarization Models in Videos over Cloud’, in Computer Science & Engineering, Visvesvaraya National Institute of Technology, Nagpur, India in March 2019. Presently, he is shouldering the responsibility as head of the Department. Dr Kumar has more than 7 years of teaching, research and administrative experience. He published more than 75 articles in reputed international conferences, book chapters and reputed journals including IEEE Transactions. His research interests include video processing, machine learning, deep learning, cloud computing, natural language processing. He is member of ACM, lifetime member of IAPR, IETE and ISTE and Senior Member of IEEE. Email: [email protected]